Detecting dissociations in single-case studies: Type I errors, statistical power and the classical versus strong...

Area of Science:

• Neuropsychology
• Cognitive Neuroscience
• Psychology

Background:

• Single-case studies are crucial for neuropsychological theory development.
• Empirical validation of dissociation criteria is essential.
• Existing methods for identifying dissociations require scrutiny.

Purpose of the Study:

• To empirically examine the Type I error rates of two methods for detecting strong dissociations.
• To evaluate the power of these methods in identifying dissociations.
• To question the practical utility of the strong versus classical dissociation distinction.

Main Methods:

• Monte Carlo simulations were employed to extend previous work on classical dissociations.
• Two methods for detecting strong dissociations were analyzed.
• Type I error rates were assessed under different misclassification definitions.

Main Results:

• When misclassifying healthy controls, Type I error rates were low for both methods.
• High Type I error rates were observed when misclassifying patients with equivalent deficits.
• The distinction between strong and classical dissociations showed questionable practical utility.

Conclusions:

• Current criteria for strong dissociations may lead to high error rates.
• The power to detect dissociations varies depending on the definition used.
• The practical value of differentiating between strong and classical dissociations is uncertain.